Software Quality - Science topic

Many companies use a SW static analysis tool, such as Klocwork, Coverity, etc., to get the SW metric. For SW quality, the defect density (how many SW defects found by the code review and/or test (unit test, integration test, etc.)) would be a leading indicator.

I am excited and curious about how your research topic will progress and the final result.

You may also want to refer to the paper below:

This paper is about source code readability measured by an Eclipse extension tool and provides insight into the readability of code for beginner programmers.

Although it is not a research that I participated in, I participated as a subject in the survey (See Section 4.2) 4 years ago and have been an undergraduate intern in this lab for about a year.

Hope this can provide some help:

These links may help you....

You may check these 2 papers

Various topics are current foucus of research in software engineering.

Some of them are followsfocus:

1. Automatic program repair

2. Effective fault localization

3. Improvement in available software metrics for integration testing and system testing.

4. Agile software development

5. Model-Driven Development

6. Software Evolution and Maintenance

7. Mining Software Repositories

8. System Verification and Validation

9. Reusable and Extendable AI models

10. ML Lifecycle management

11. ML model Catalogs and Searching for Rapid Development

12. Testing, Correctness, Debugging & Interpretability of AI/ML applications

Hello;

I suggest the tools may help you:

1- Selenium (https://www.selenium.dev/).

2- Junit (junit.org)

3- reactjs (https://reactjs.org/)

Industry would usually go for compliance of statute. In case anybody is really interested on a protocol, he would like to follow international best practices. Some of them will listen to what their consultants say.

Do you have an example you can share? Thanks in advance.

You can use any related research in which went through with CFA (confirmatory factor analysis). Moreover, those research should show good psychometric properties as well. You may adopt or adapt the items based on your research purposes.

I agree with Mazen like Ghareb the ISO/IEC 25000 is the last Framework for Software Quality Measurement.

Big Data database systems can be very diverse, because they can be different in virtually every respect. In individual institutions and companies in which Big Data database systems are built and developed, they may differ, among others: 1. Technical and technological characteristics of servers, disks and other elements of IT and network equipment, potential of data processing speed. 2. The type and generation of operational database systems supporting specific Big Data databases. 3. Installed applications for archiving, indexing, searching and analyzing information collected in database systems. 4. The possibilities of expanding and improving specific, existing Big Data database systems. Therefore, if there is a need to accurately compare specific Big Data database systems functioning in institutions and companies, then a universal, multi-faceted, multi-factor scoring model should be built, thanks to which it will be possible to conduct such comparative analyzes.

Best wishes

Turnitin is a good one for dictating plagiarized content.

I believe for estimating the cost estimation models, there are numerous validation/accuracy evaluation techniques such as Mean of Relative Errors (MRE) [43], Pred (0.25) and the median magnitude of relative error (MDMA) are the most used by the researchers in the field.

Accuracy Comparison of Analogy-Based Software Development Effort Estimation Techniques by Ali Idri might help you to understand.

Following readings may help:

1. M. Shepperd and C. Schofield, "Estimating software project effort using analogies," IEEE Transactions on Software Engineering, vol.13, pp. 736-743, 1997.

2. S.D Conte, HE Dunsmore, and VY Shen, Software Engineering Metrics and Models, Benjamin-Cummings, Menlo Park CA, 1986.

Hi,

One way to "connect" inheritance with software quality are metrics. MOOD is a good starting point, e.g. with the

Method Inheritance Factor (MIF).

Here is an intro https://www.google.com/url?sa=t&source=web&rct=j&url=http://www8.cs.umu.se/education/examina/Rapporter/MuktamyeSarker.pdf&ved=2ahUKEwiFk_j09cThAhWVdHAKHQX0B6YQFjABegQIDhAF&usg=AOvVaw3wcCXA3uPoxngYDz5KRCGr. There are many more paper on MOOD and OO metrics.

Regards

Roland

I introduce this paper for evaluating uncertainty factors maybe affect software reliability.

THE MAJOR UNCERTAINTY FACTORS AFFECTING IN

SOFTWARE COST ESTIMATION, G Rajkumar et al , Int.J.Computer Technology & Applications,Vol 4 (3),419-424.

Extrinsic Influence Factors in Software Reliability:

A Study of 200,000 Windows Machines

Christian Bird, Venkatesh-Prasad Ranganath, Thomas Zimmermann, Nachiappan Nagappan, Andreas Zeller.

I would also recommend journals that support the journal-first model, where you could present your publications at prestigious conferences, such as ICSE and ESEC/FSE:

No such tools are used for quality requirements ...but some methods are used..

Thanks @ Wiem Khlif.

Are these standards available free online? My work concern with quality attributes such as security, usability, flexibility standard such as ISO 9126. Any quality standards available free online for research purpose?

i suggest you to read this paper:

wiem

Thank you all for your contributions!

However, every implemented performance evaluation measure has advantages and disadvantages (Powers 2011, Abran 2010). That is why (in my opinion) none of these evaluation measures won a global consensus among software researchers and practitioners. 

Is it possible to start looking for a way to combine more than one evaluation measure together to evaluate prediction models performance? For example, Area Under ROC Curve combines two performance evaluation measures together, namely True Positive Rate and False Positive Rate. Is there a way to combine others and even possibly more than two or three, etc. Thank you.

Hussam

I would suggest "The Art of Software Testing" by GJ Myers. 2. Secure Coding Principles by Robert Seacord

You can also try the following links for downloading each volume:

Check rule 5 again. What is source of these rules?

Hi Muhammad, 

  You can have a look at open source C projects in gitHub. 

As James said, everyone is responsible, whether they know it or not (i.e. manager imposing unrealistic deadlines). Another interesting small book discussing these quality issues is: "Facts and Fallacies of Software Engineering" from Robert Glass.

Dear Faheem Haider,

One possible way to start is to choose a focus between "process" and "product". One can address Quality Assurrance (QA) either by a process perspective or by a product perspective. Task and activities for QA, best practices, standards are examples of process related topics you can address. Also, you can address product metrics regarding quality assurance goals. Anyway, it is a broad field. Particularly, I would suggest "code review" as part of the quality assurrance process a nice topic to go deep into. In the "product perspective", I suggest research on "code clones" as many recent works have found an association between code clones and software quality.   

In an absolute since, the relationship between quality and software is a function of the marketplace and the cost of re-deployment. For example, software built for in-house application can be of lower quality than software controlling an implanted defibrillator, because the cost of deploying a defective defibrillator is high. For less critical applications the primary driver is the cost of re-deployment; which consist of the cost of repair plus the cost of redistribution. For example, when embedded application were distributed on various roms, the cost to repair was low with a very high cost to re-distribute.

I'd like to know what you mean by "quality" and "after compiling" before fully answering. If you mean conformance to requirements, the only way you can improve a compiled version of the software is to wrap it in other software that disables some of its functionality (e.g., by limiting ranges on parameter values, controlling the sequence in which APIs are invoked, or excising some APIs altogether). This doesn't change the requirements conformance of the software component itself but may improve the quality of the system in which it is embedded. The extreme case would be to take a bad module and to put it together with another module that completely replaces its functionality. Though the software is part of the deliverable it is never called, its function having been taken over by the surrogate.

If you embed that software into software that takes none of these approaches, you can only *lower* the overall quality. Think of a hardware analogy: if a board has just one badly designed integrated circuit that has a certain number of failures in a trillion (FITs), adding more chips will only cause their FIT values to lower the end-to-end reliability — unless a high-reliability component is used to hide/wrap the faults of the one under consideration.

GPSRs are challanging- GPSR made simple will be great to use

Best wishes

I like your idea and look forward to sharing 

You could use "SonarQube" : A tool that is dedicated to gather many of the important software metrics for Continuous Inspection : http://www.sonarqube.org/. I highly recommend the books : "SonarQube in Action" by G. Ann Campbell and Patroklos P. Papapetrou and "Software in Zahlen, Die Vermessung von Applikationen" by Harry M. Sneed, Richard Seidl and Manfred Baumgartner (in German language).

Yes Prof.David C.Rine same interest area.

Our goal is to predict the classes that will contain faults at the next release of an Object

Oriented System. You know sir, current software metrics and defect data are used to

construct the prediction model for the next release of software. Lots of the Ds (datasets) which locate in NASA PROMISE repository. We had used NASA datasets in PROMISE repository as KC1 (Chidamber-Kemerer Object-Oriented) and method level metrics such as Halstead and McCabe. But NASA datasets are known by everyone.

We want to create a new data set and we has lots of source code (millitary,commercial,education)

I need a software to find object oriented metrics from our softwares. (I know McCabe software but so expensive) We will use these data to create classmetricInfo table for Weka software. 

Thank you so much Prof.David C.Rine for your interest and recommendation, I download i like them all

Hi,

There are many models, below is a sample

1- ISO standard models

There are many International Standard Organization (ISO) standard frameworks that can support software quality; however they are generally checking list guides. For example, the ISO 9000 family has been criticized not to be used for software quality (Stelzer et al. 1997). The ANSI/IEEE 730-2002 of Poston (1985) support quality assurance plans. Capability Maturity Model Integration (CMMI) of Ahern et al. (2008) is a process improvement training and certification program. So, these ISO standards are neither designed to address quality-in-use nor specific product characteristics.

Recently, the Software Product Quality Requirements and Evaluation (SQuaRE) ISO standards’ series has been derived from the blending of ISO/IEC 9126 and ISO/IEC 14598 series of standards. The purpose of the SQuaRE series of standards is to assist in developing and acquiring software products with the specification of quality requirements and evaluation. From the viewpoint of the stakeholders the quality requirements are specified, while the quality of the product is evaluated based on this specification. The evaluation is based on the chosen quality model, quality measurement and quality management process.

The ISO/IEC 25010:2011 standard (ISO 25010 hereafter), a part of the SQuaRE series, has two major dimensions; Quality-in-Use and Product Quality. The former specifies characteristics related to the human interaction with the system and the latter specifies characteristics intrinsic to the product. It also has one complementary data model (the ISO/IEC 25012).

2- customized models

La and Kim (2013) adopted the effectiveness and productivity of ISO 9126 model for a service-based mobile ecosystem. Based on the ISO 9126 model structure, Osman and Osman (2013) proposed three categories to model quality-in-use of mobile government systems. The categories are usability, acceptance, and user experience. The quality-in-use were calculated with a defined set of tasks and a post testpost-test of a questionnaire. Oliveira, Tereso, and Machado (2014) linked the ISO 9126 characteristics; functionality, usability, and operability to a set of criteria/requirement of project management tools in order to help project managers evaluate project management tools.

Quality, Quality-in-Use, Usability and User experience (2Q2U) model (Lew et al. 2010) extends the ISO 25010 model with information quality, learnability-in-us, actual usability, and user experience characteristics. Semantic Web Exploration Tools Quality-in-Use Model (SWET-QUM) model (González et al. 2012) extends the ISO 25010 with several metrics related to Semantic Web Exploration tools. In SWET-QUM tasks are customized to web exploration tools, and are manually linked to the ISO QinU model.

Issa

heck this url mybe it's help you

Regards,

Configuration Management Repositories are based on files, so there is no way around that - especially when you want to track change history.  Nothing precludes you from having an enforceable policy of object definitions in .h files and methods in the .c files (as an example).

Ada had the specification part and the body part when implementing objects and their associated methods.

Newer implementations via XML (and other) may allow for single file use for object definition and method implementation, but some other parser would have to be able to separate out the parts.  Again, still have a file.

That is a tricky question and highly depends on the prediction model. Many companies, such as Microsoft, use prediction models to assess quality risk and potential defective areas, for example to guide software testing. The same can be true for agile processes too. If you want to reduce test effort, you can use prediction models based on process measures that detect process issues that have shown to correlate with defects. Use the attaches change burst paper as an example.

There are two important factors about prediction models

1) The model itself is just an indicator and should not be trusted entirely. It is good practice to compare the models output with engineer expectations. Usually, you see a large overlap between both risk lists. The difference between both lists is the interesting parts--areas that the model considers risky while the engineers do not have this area on the radar. These might be the areas of highest risk as nobody cared about them. This can be true for both static waterfall, agile, and intermediate processes.

2) The model should be tightly bound against the development period and you should ensure that already fixed bugs are recognized. Most prediction models published do not consider continuous usage of these models and are based on measurements that cannot capture already applied bug fixes. For example, consider a prediction model based on code dependencies. Now assume, you hunt all the bugs in all the files predicted most defect-prone and you run the model again. Well, you will get the same result from the model, because fixing a bug will not change code dependencies significantly. At this point, your metric is gone since you deliberately destroyed the correlation between dependencies and defects by fixing the defects. Such a model will not work in an agile setting. There it is important to choose dynamic and self-adapting measurements that capture already fixed bugs. An example might be change bursts, test metrics (e.g. increase in coverage, number of new scenario tests), and in general measurements that capture the difference between two revisions.

Thank you so much I closed it

Dear Nawel, I'm writing a tool for automatic building of Goal models from NL requirement documents. I found DODT very interesting, but I can't find anything else except that document you posted here: can you suggest me were to download the tool or the source code (or any other document)? Many thanks!

Probably because a smart phone is seen as "yet another input device"?

I think Promise is a good start so . I think there are no "universally" valid prediction model for faults, so many variables involved (from structural metrics to complexity of the task at hand)